Modelling hydrology and water quality in the pre-alpine/alpine Thur watershed using SWAT

Summary

In a national effort, since 1972, the Swiss Government started the “National Long-term Monitoring of Swiss Rivers” (NADUF) program aimed at evaluating the chemical and physical states of major rivers leaving Swiss political boundaries. The established monitoring network of 19 sampling stations included locations on all major rivers of Switzerland. This study complements the monitoring program and aims to model one of the program’s catchments – Thur River basin (area 1700 km²), which is located in the north-east of Switzerland and is a direct tributary to the Rhine. The program SWAT (Soil and Water Assessment Tool) was used to simulate all related processes affecting water quantity, sediment, and nutrient loads in the catchment. The main objectives were to test the performance of SWAT and the feasibility of using this model as a simulator of flow and transport processes at a watershed scale. Model calibration and uncertainty analysis were performed with SUFI-2 (Sequential Uncertainty FItting Ver. 2), which was interfaced with SWAT using the generic iSWAT program. Two measures were used to assess the goodness of calibration: (1) the percentage of data bracketed by the 95% prediction uncertainty calculated at the 2.5 and 97.5 percentiles of the cumulative distribution of the simulated variables, and (2) the d-factor, which is the ratio of the average distance between the above percentiles and the standard deviation of the corresponding measured variable. These statistics showed excellent results for discharge and nitrate and quite good results for sediment and total phosphorous. We concluded that: in watersheds similar to Thur – with good data quality and availability and relatively small model uncertainty – it is feasible to use SWAT as a flow and transport simulator. This is a precursor for watershed management studies.

Introduction

During the last three decades, in Switzerland, as well as in other European countries, extensive and costly measures have been taken to reduce pollution input by point sources. The measures included installation of advanced waste water treatment and regulations for restricted use of phosphorous (phosphate ban in household detergents) and toxic substances. In 1972, the “National Long-term Monitoring of Swiss Rivers” (NADUF) program was initiated as a cooperative project between what is now the Swiss Federal Office for the Environment (FOEN), and the Swiss Federal Institute of Aquatic Science and Technology (Eawag) (Binderheim-Bankay et al., 2000) (www.naduf.ch). The program monitored the chemical and physical states of Swiss rivers to evaluate the effectiveness of water protection measures undertaken by Swiss national environmental protection agencies. Several NADUF catchments and stations, including the one investigated in this study – the pre-alpine/alpine Thur River basin, which represents a low level of anthropogenic pollution – serve as reference stations for other international water monitoring programs, e.g., International Commission for the Protection of the Rhine (IKSR), the Global Environmental Monitoring System – United Nations Environmental Program/World Health Organization (GEMS-UNEP/WHO) (Jakob et al., 2002). After changing the water protection law in the late 1980s to early 1990s, the positive effects of these measures were reported at several monitoring stations. In the Swiss part of the Rhine watershed, the international target of 50% reduction in the total inputs into surface waters of P and N was achieved for P (reduction of 51%) but not for N (reduction of 23%) (Prasuhn and Sieber, 2005). Also, lead concentration decreased by 80–90% during the same time period. Furthermore, with the adoption of a new “ecologically oriented” agricultural management in 1993, which included animal friendly farming, balanced use of fertilizers, appropriate proportions of ecological compensation areas, suitable crop rotation, soil erosion protection, and measured use of pesticides – decreasing trends of nutrients in Swiss water bodies were reported as well (Jakob et al., 2002, SAEFL, 2002). The total phosphorous and nitrogen concentrations decreased significantly by 28% and 14%, respectively, from 1985 to 2001 (Prasuhn and Sieber, 2005). However, the problem of non-point source pollution still exists and is associated primarily with the agricultural applications of mineral (ammonium, nitrate) and organic (liquid and solid manure) fertilizers. It should be noted that the landuse change during the period of 1980–1995 has been quite insignificant in the Thur region as indicated by the first (from 1979 to 1985) and the second (from 1992 to 1995) landuse maps complied by the Swiss Federal Statistical Office (www.bfs.admin.ch). The latest landuse map shown in Fig. 1 indicates a predominantly agricultural region.

Surface runoff, especially immediately after a storm, is an important medium of transport for non-point source pollution. Runoff from different landuses may be enriched with different kinds of contaminants. For example, runoff from agricultural lands is generally enriched with sediments, nutrients and pesticides, whereas runoff from actively developed urban areas contains heavy metals, hydrocarbons, chloride and other contaminants (Huber, 1993). Due to the significant reduction in the loads from point sources in the past years, the relative significance of diffuse sources of pollution in Swiss waters has increased. Presently in Switzerland, wash-out and runoff from agricultural lands contributes to a greater extent to the impairment of natural waters than it was a few decades ago (Prasuhn and Sieber, 2005).

Inverse modelling (IM) has in recent years become a very popular method for calibration (e.g., Beven and Binley, 1992, Abbaspour et al., 1997, Simunek et al., 1999, Duan et al., 2003, Gupta et al., 2003, Wang et al., 2003). IM is concerned with the problem of making inferences about physical systems from measured output variables of the model (e.g., river discharge, sediment concentration). This is attractive because direct measurement of parameters describing the physical system is time consuming, costly, tedious, and often has limited applicability. Because nearly all measurements are subject to some uncertainty, the inferences are usually statistical in nature. Furthermore, because one can only measure a limited number of (noisy) data and because physical systems are usually modelled by continuum equations, no hydrological inverse problem is really uniquely solvable. In other words, if there is a single model that fits the measurements there will be many of them. Our goal in inverse modelling is then to characterize the set of models, mainly through assigning distributions (uncertainties) to the parameters that fit the data and satisfy our presumptions as well as other prior information.

To make the parameter inferences quantitative, one must consider: (1) the error in the measured data (driving variables such as rainfall and temperature), (2) the error in the measured output variables (e.g., river discharges and sediment concentrations used for calibration), and (3) the error in the conceptual mode (inclusion of all the physics in the model that contributes significantly to the data).

The objective of this research study was to evaluate the application of a mechanistic modelling approach as a complementary technique to the monitoring program in investigating the relative impacts of different types of landuse and agricultural managements on water quality and quantity of the Thur River. A number of simulators such as SWAT (Soil Water Assessment Tool) (Arnold et al., 1998), HSPF (Hydrologic Simulation Program Fortran) (Bicknell et al., 1996), and SHETRAN (Ewen et al., 2000) could have been used in this study. Several comparisons of these models indicated similarly reasonable results in simulating discharge, phosphorous, and sediment (e.g., Singh et al., 2005, Borah and Bera, 2004). We chose SWAT because of its availability and user-friendliness in handling input data. SWAT was evaluated by performing calibration and uncertainty analysis using SUFI-2 (sequential uncertainty fitting ver. 2) algorithm (Abbaspour et al., 2004), which is a semi-automated inverse modelling procedure for a combined calibration-uncertainty analysis. The available time series data on discharge, sediment, nitrate, and total phosphorus loads at the watershed outlet as well as some constraints on sediment and nutrients from different landuses were used to perform calibration and validation studies.